Ethernet QoS Configuration Guide - Avanzada 7 · 2 E-DOC-CTC-20080229-0002 v1.0 About this Ethernet...

Po

wer

Eth

ern

et

WLA

N

Plu

g-i

n

ISD

N

Inte

rnet

DS

L

Ethernet QoS Configuration GuideR7.4 and higher

Thomson Gateway

Thomson GatewayEthernet QoS Configuration GuideR7.4 and higher

Copyright

Copyright ©1999-2008 Thomson. All rights reserved.

Distribution and copying of this document, use and communication of its contents is not permitted without written authorization from Thomson. The content of this document is furnished for informational use only, may be subject to change without notice, and should not be construed as a commitment by Thomson. Thomson assumes no responsibility or liability for any errors or inaccuracies that may appear in this document.

Thomson Telecom BelgiumPrins Boudewijnlaan, 47 B-2650 Edegem Belgium

http://www.thomson-broadband.com

Trademarks

The following trademarks may be used in this document:

DECT is a trademark of ETSI.

Bluetooth® word mark and logos are owned by the Bluetooth SIG, Inc.

Ethernet™ is a trademark of Xerox Corporation.

Wi-Fi® and the Wi-Fi logo are registered trademarks of the Wi-Fi Alliance. "Wi-Fi CERTIFIED", "Wi-Fi ZONE", "Wi-Fi Alli-ance", their respective logos and "Wi-Fi Protected Access" are trademarks of the Wi-Fi Alliance.

UPnP™ is a certification mark of the UPnP™ Implementers Corporation.

Microsoft®, MS-DOS®, Windows®, Windows NT® and Windows Vista® are either registered trademarks or trademarks of Microsoft Corporation in the United States and/or other countries.

Apple® and Mac OS® are registered trademarks of Apple Computer, Incorporated, registered in the United States and other countries.

UNIX® is a registered trademark of UNIX System Laboratories, Incorporated.

Adobe®, the Adobe logo, Acrobat and Acrobat Reader are trademarks or registered trademarks of Adobe Systems, Incor-porated, registered in the United States and/or other countries.

Other brands and product names may be trademarks or registered trademarks of their respective holders.

Document Information

Status: v1.0 (March 2008)Reference: E-DOC-CTC-20080229-0002Short Title: Config Guide: Ethernet QoS R7.4 and higher


Contents

About this Ethernet QoS Configuration Guide........................ 1

1 Introduction.................................................................................. 3

2 Ethernet QoS Overview .............................................................. 5

3 QoS Capabilities of the Ethernet Switch................................... 7

3.1 Classification and Queuing ............................................................................. 8

3.2 Rate Control .................................................................................................. 11

3.3 Storm Control ............................................................................................... 15

4 QoS Capabilities of the Ethernet Bridge ................................. 17

4.1 802.1p User Priority Regeneration................................................................ 18

4.2 Ingress Classification.................................................................................... 20

4.2.1 Medium Priority.................................................................................................................................... 21

4.2.2 Evaluated Priority ................................................................................................................................. 22

4.2.3 Internal Priority ..................................................................................................................................... 26

4.3 Queuing......................................................................................................... 28

4.4 Priority Tagging of Egress Traffic................................................................. 31

4.5 802.1p User Priority Preservation................................................................. 32

5 Logical Ethernet Interfaces and QoS ....................................... 33

6 Layer 2 QoS and Layer 3 QoS Interaction ............................... 35

E-DOC-CTC-20080229-0002 v1.0 i

ii

Contents

E-DOC-CTC-20080229-0002 v1.0

About this Ethernet QoS Configuration Guide


Used Symbols

TerminologyGenerally, the Thomson Gateway356 will be referred to as Wireless USB Adaptor in this Ethernet QoS Configuration Guide.

Typographical ConventionsFollowing typographical convention is used throughout this manual:

Sample text indicates a hyperlink to a Web site.

Example: For more information, visit us at www.thomson-broadband.com.

Sample text indicates an internal cross-reference.

Example: If you want to know more about guide, see “1 Introduction” on page 7”.

Sample text indicates an important content-related word.

Example: To enter the network, you must authenticate yourself.

Sample text indicates a GUI element (commands on menus and buttons, dialog box elements, file names, paths and folders).

Example: On the File menu, click Open to open a file.

Documentation and software updatesThomson continuously develops new solutions, but is also committed to improving its existing products.

For more information on Thomson's latest technological innovations, documents and software releases, visit us at http://www.thomson-broadband.com.

A note provides additional information about a topic.

A caution warns you about potential problems or specific precautions that need to be taken.

E-DOC-CTC-20080229-0002 v1.0 1



2


OverviewFirst, this Ethernet QoS Configuration Guide shortly lists the Ethernet QoS features of the Thomson Gateway and the relevant standards. Some background information on several Ethernet QoS-related concepts is also provided. Next, this document describes the configuration of each Ethernet QoS feature on the Thomson Gateway in detail. CLI (Command Line Interface) commands are used for the configuration.

This document is structured as follows:

Topic Page

“1 Introduction” 3

“2 Ethernet QoS Overview” 5

“3 QoS Capabilities of the Ethernet Switch” 7

“4 QoS Capabilities of the Ethernet Bridge” 17

“5 Logical Ethernet Interfaces and QoS” 33

“6 Layer 2 QoS and Layer 3 QoS Interaction” 35

E-DOC-CTC-20080229-0002 v1.0

1| Introduction

1 Introduction

IntroductionEthernet QoS, or Layer 2 QoS, is the subject of this document. In this chapter, we shortly list the Ethernet QoS features that are supported by the Thomson Gateway, the relevant standards and related documents.

Supported Ethernet QoS featuresThomson Gateway devices support following Ethernet QoS features:

QoS features of the Ethernet switch

Classification

Queuing and scheduling

Pause frame flow control

Rate control (ingress rate limiting and egress traffic shaping)

Storm control

QoS features of the Ethernet bridge

802.1p user priority regeneration

Ingress classification

Queuing

Priority tagging

802.1p user priority preservation

QoS features of the logical Ethernet interfaces

WAN-side priority tagging

Standards compliancyThomson Gateway devices are compliant with following standards, which are relevant to Ethernet, Ethernet bridging, VLAN bridging, VLAN user priorities and stacked VLANs:

IEEE Std 802.3 - 2000: Part 3: Carrier sense multiple access with collision detection (CSMA/CD) access method and physical layer specifications.

IEEE Std 802.1D - 2004: IEEE standard for local and metropolitan area networks - Media Access Control (MAC) bridges.

IEEE Std 802.1Q - 1998: IEEE standard for local and metropolitan area networks - Virtual bridged local area networks.

IEEE Std 802.1p: Traffic Class Expediting and Dynamic Multicast Filtering. This standard is merged into 802.1D-2004.

IEEE Std 802.1ad - 2005: IEEE standard for local and metropolitan area networks - Virtual bridged local area networks - Amendment 4: Provider bridges.

Related documentsOther Layer 2 related features, such as Ethernet and VLANs, are described in other documents. See the “Ethernet Configuration Guide” and the “VLAN Configuration Guide” for more information.

E-DOC-CTC-20080229-0002 v1.0 3

4

1| Introduction

E-DOC-CTC-20080229-0002 v1.0

2| Ethernet QoS Overview

2 Ethernet QoS Overview

IntroductionThe overall QoS framework of Thomson Gateway devices consists of several QoS elements to support demanding applications such as Voice over IP (VoIP), IPTV, management and so on. The two QoS elements within the scope of this document are the Ethernet switch and the Ethernet bridge.

Ethernet QoS functionalityThe Ethernet QoS functionality mainly consists of two steps:

At the ingress side, the incoming frames are classified. This means that an internal class, i.e. priority, is assigned to each frame.

At the egress side, each frame is queued into the correct queue, based on the assigned internal class.

Internal class informationThe internal class of a frame or packet plays an important role during both classification and queuing. There are 16 internal classes, numbered from 0 through 15.

All QoS elements that are part of the overall QoS framework work together and share the internal class information for each frame or packet. This means that when classification is done by the Ethernet bridge, the resulting internal class can be used or overruled by the IP router. The other way around, when classification is done by the IP router, the resulting internal class can be used to bridge the traffic to the correct ATM PVC or the correct QoS queue of a single ATM PVC.

Classification-related frame fieldsIn order to assign an internal class to a received frame, several information fields of the frame can be used. Following information fields will be used in this document:

802.1p user priority field: this is a three-bit field in the VLAN tag of an Ethernet frame, indicated in red in following illustration:

DestinationMAC Address

SourceMAC Address

Length/Type

TPID TCI FCS

6 6 2 42 2

PadData/LLC

0 .. 1500 46 .. 0 bytes

Min. 46 bytes - Max. 1500 bytes

Min. 68 bytes - Max. 1522 bytes

Priority CFI VID

3 1 12 bits

VLAN tag

E-DOC-CTC-20080229-0002 v1.0 5

6

2| Ethernet QoS Overview

IP precedence field: this is a three-bit field in the header of an IP packet. It is part of the ToS (Type of Service) byte of an IPv4 header or the Traffic Class byte of an IPv6 header.

DSCP (Differentiated Services Code Point) field : this is a six-bit field in the header of an IP packet. It is part of the Differentiated Services (DS) field of the IP header. The DS field replaces the old ToS byte of the IPv4 header and the old Traffic Class byte of the IPv6 header. The values of the DSCP field provide limited backwards compatibility with the IP precedence field.

Precedence

8 bits

DelayThrough-

putReliability ReservedReserved

IP Precedence Field

1 2 3 4 5 6 70

Differentiated Services Code Point Field

8 bits

Explicit CongestionNotification

Differentiated Services Code Point (DSCP)

1 2 3 4 5 6 70

E-DOC-CTC-20080229-0002 v1.0

3| QoS Capabilities of the Ethernet Switch

3 QoS Capabilities of the Ethernet Switch

IntroductionWhich QoS features of the Ethernet switch are available depend on the used Thomson Gateway device. We make a distinction between residential devices and business devices. Whenever necessary, it is noted to which type of devices the feature is applicable.

Listing the QoS capabilities of the switchTo display the QoS capabilities of the Ethernet switch, execute the command :eth switch info.

To list the QoS capabilities of a residential device, for example the Thomson ST780, execute following command:

To list the QoS capabilities of a business device, for example the Thomson ST620, execute following command:

OverviewFollowing QoS features are described in this chapter:

=>:eth switch infoEthernet switch capabilities:+ Number of external ports=4+ QoS supported: number of queues=4

+ diffserv classifier supported+ Storm control supported, possible values: 100 200 300 400

=>:eth switch infoEthernet switch capabilities:+ Number of external ports=4+ QoS supported: number of queues=4

+ realtime queue supported+ diffserv classifier supported+ tos classifier supported

+ Ingress shaping supported: number of ingress shapers=2+ Egress shaping supported: number of egress shapers=1+ Storm control supported, possible values: 33 50 100 200

Topic Page

“3.1 Classification and Queuing” 8

“3.2 Rate Control” 11

“3.3 Storm Control” 15

E-DOC-CTC-20080229-0002 v1.0 7

8


3.1 Classification and Queuing

IntroductionSimply explained, classification and queuing work as follows:

At the ingress side, each frame is classified: the Ethernet switch assigns a priority to each received frame, based on the configured classification criterion.

At the egress side the frame is queued: the Ethernet switch uses the priority of the frame to map the frame onto the correct queue. The number of queues can be configured (up to four).

For the mapping from the classification result (priority) to the correct queue, the Ethernet switch has internal mapping tables. You cannot configure these mapping tables.

The CLI commands described in this section belong to the command group :eth switch qos.

Listing classification and queuing informationTo obtain information on QoS support, the configured classification criterion and queues, execute the command :eth switch qos list.

By default, QoS support is disabled:

If QoS support is enabled, the listed information consists of three parts, as shown in the following illustration:

Enabling QoS supportThe common QoS parameters can be configured with the command :eth switch qos config.

Following parameter must be specified:

State: this parameter is used to enable or disable QoS support on the Ethernet switch. By default, QoS support is disabled.

To disable QoS support, set this parameter to disabled.

To enable QoS support, set this parameter to enabled. In addition, define at least two queues per switch port.

=>:eth switch qos listNo queues configured => QoS disabled.

:eth switch qos listQoS enabled on switch. 2 queues configured. Realtime disabled.

Queue Weight0 331 67

Port Classifier FlowControl1 802.1P off2 802.1P off3 802.1P off4 802.1P off

Enabling QoS supportcommand config

Classificationcommand ifconfig

Queuingcommand weights

E-DOC-CTC-20080229-0002 v1.0


Optionally, two other parameters can be specified:

NbrOfQueues: this parameter indicates the number of queues per switch port. This is a number within the range from 0 through 4.

Realtime: the highest priority queue can be set as a real-time queue by enabling this parameter. By default, this parameter is disabled.

ClassificationClassification can be configured for each switch port with the command :eth switch qos ifconfig.

Two parameters must be specified:

Port: this parameter specifies the switch port to be configured, indicated by a number from 1 through 4.

Mode: this parameter indicates the used classification criterion.

None: all traffic will be classified as low priority.

High: all traffic will be classified as high priority.

802.1p: classification is based on the value of the 802.1p user priority field in the VLAN tag of a VLAN-tagged or priority-tagged frame. This is the default classification criterion.

Diffserv: classification is based on the value of the DSCP field in the IP header of a packet.

ToS: classification is based on the value of the IP precedence field in the IP header of a packet.

Optionally, following parameter can be specified:

Flowcontrol: this parameter is used to enable or disable pause frame flow control. By default, flow control is disabled.

A flow control mechanism can be used when the data rate sent by a device exceeds the input data rate of the receiver. In this case, the receiver sends a pause frame which stops the transmission of the sender for a specified period of time. A pause frame is a MAC control frame.

At least two queues must be defined in order to enable QoS support on the Ethernet switch.

In case of a residential device, the parameter realtime is not applicable.

=>:eth switch qos config state=enabled nbrOfQueues=3 realtime=enabled=>:eth switch qos listQoS enabled on switch. 3 queues configured. Realtime enabled.

Queue Weight0 141 282 58

Port Classifier FlowControl1 802.1P off2 802.1P off3 802.1P off4 802.1P off

At the egress switch ports, the classification criterion can be set to none as all classification is done at the ingress side.

In case of a residential device, the ToS mode is not supported.

E-DOC-CTC-20080229-0002 v1.0 9

10


For example, configure the classification as follows:

QueuingAt the egress side, frames are queued. The used scheduling algorithm is WFQ (Weighted Fair Queuing). This means that the queues are being served based upon their weight. A queue with a weight of W% will achieve an average data rate of W% of the link data rate. The higher the weight, the higher the priority.

The weight of each queue can be configured with the command :eth switch qos weights.

For each queue, the weight must be specified:

QueueX: this parameter indicates the weight of queue X in WFQ, expressed in percent. The weight can range from 0% through 100%. By default, the weights are as follows:

2 queues: 33% and 67%.

3 queues: 14%, 28% and 58%.

4 queues: 6%, 13%, 27% and 54%.

=>:eth switch qos ifconfig port=2 mode=high flowcontrol=enabled=>:eth switch qos listQoS enabled on switch. 3 queues configured. Realtime enabled.

Queue Weight0 141 282 58

Port Classifier FlowControl

1 802.1P off

2 High on

3 802.1P off

4 802.1P off

QoS support must be enabled on the Ethernet switch before the classification parameters can be configured.

=>:eth switch qos weights queue0=24 queue1=31 queue2=45=>:eth switch qos listQoS enabled on switch. 3 queues configured. Realtime enabled.

Queue Weight

0 24

1 31

2 45

Port Classifier FlowControl1 802.1P off2 High on3 802.1P off4 802.1P off

The sum of the weights of the queues must be equal to 100.

QoS support must be enabled on the Ethernet switch and the correct number of queues must be defined before the weights can be configured.

E-DOC-CTC-20080229-0002 v1.0


3.2 Rate Control

IntroductionRate control is used to meter and limit the rate of a data stream. Rate control can be performed through rate limiting or shaping:

Ingress rate limiting

At the ingress side, rate limiting is used to restrict the bit rate of specific traffic types. Excess traffic is dropped.

Two ingress traffic shapers, also referred to as buckets, can be defined. Specific traffic types can be assigned to one of the two traffic shapers. Per switch port, a speed and burst size can be specified.

Egress shaping

At the egress side, traffic shaping is used: traffic is buffered and then sent at a predefined rate. In this case, excess traffic will not be dropped as long as the buffer is not congested.

Only one shaper or bucket is used. Thus, all egress traffic must pass the egress rate control checking. Per switch port, a speed and burst size can be specified.

The CLI commands described in this section belong to the command group :eth switch shaper.

How does a token bucket algorithm work?Rate control is typically achieved using one or more token bucket algorithms. The operation of such a token bucket algorithm can be explained as follows:

The bucket or traffic shaper is characterized as follows:

Tokens are added to the bucket at a given speed.

The bucket can contain only a limited number of tokens, defined by the size of the bucket.

Every time a frame arrives at the bucket, a token is taken from the bucket before the frame is accepted and forwarded.

If a frame arrives at the bucket and no tokens are available (the bucket is empty), the frame is dropped.

As a result, the speed tokens are added to the bucket limits the rate of the data frames and the size of the bucket defines the maximum burst size of the data stream. Both the speed limit and burst size are configurable.

Rate limiting and shaping are not supported by residential devices. The description in this section applies to business devices.

E-DOC-CTC-20080229-0002 v1.0 11

12


Listing the rate control configurationTo look up the rate control configuration, execute following command:

Configuring the ingress traffic shapersTwo different ingress traffic shapers can be defined with the command :eth switch shaper config. These two traffic shapers exist for each switch port and can not be configured per port.

One parameter must be specified:

Shaper: this parameter specifies the traffic shaper instance to be configured. Only two traffic shaper instances exist, indicated with a number (0 or 1).

Optionally, different parameters can be specified to configure the selected traffic shaper instance.

The traffic types to be shaped must be specified. For each traffic type, the corresponding parameter can be enabled or disabled. By default, all traffic types are disabled.

Unicast: unicast traffic

Multicast: multicast traffic

Broadcast: broadcast traffic

Control: MAC control traffic

Unknown: traffic that resulted in a destination look up failure (DLF). This means that the destination address is unknown.

=>:eth switch shaper iflistPort Direction Shaper State Speed BurstSize Flags1 egress 0 off 0 0 nnnnnn1 ingress 0 off 0 0 ------1 ingress 1 off 0 0 ------2 egress 0 off 0 0 nnnnnn2 ingress 0 off 0 0 ------2 ingress 1 off 0 0 ------3 egress 0 off 0 0 nnnnnn3 ingress 0 off 0 0 ------3 ingress 1 off 0 0 ------4 egress 0 off 0 0 nnnnnn4 ingress 0 off 0 0 ------4 ingress 1 off 0 0 ------State: configurable per port and shaper type (ingress or egress).Speed & BurstSize: configurable per port and shaper. Only limited set of values.

Input is rounded.Flags: configurable per ingress shaper: u(nicast), m(ulticast), b(roadcast),

c(ontrol), a(ddress unknown), d(iscard), n(ot applicable).The same frametype can not be set on different ingress shapers.

You can not enable the same traffic type for both traffic shaper instances.

E-DOC-CTC-20080229-0002 v1.0


The way excess traffic is handled must be specified:

Discard: if this parameter is enabled, excess frames are discarded. If the parameter is disabled, flow control is used (if available) and a pause frame is sent out. By default, this parameter is disabled.

The egress traffic shaperOnly one egress traffic shaper exists for each switch port. This egress traffic shaper can not be configured:

All traffic types must pass through the traffic shaper.

Excess traffic is not discarded as long as the queue that contains this traffic is not congested.

Configuring per port rate controlRate control can be configured on a per port basis. To configure rate control on a specific switch port, execute the command :eth switch shaper ifconfig.

Per port and per traffic direction the traffic shaper can be selected. Therefore, three parameters must be specified:


Ingress: this parameter is used to select the traffic direction.

If the parameter is enabled, the ingress traffic direction is selected.

If the parameter is disabled, the egress traffic direction is selected.

Shaper: this parameter selects the (ingress) traffic shaper instance. Only two ingress traffic shaper instances exist, indicated with a number (0 or 1). If the egress traffic direction is selected, the value of this parameter is not relevant and set to 0.

Once the combination of port, traffic direction and traffic shaper is selected, several optional parameters can be set:

State: this parameter specifies whether rate control is used or not. By default, rate control is disabled. For the ingress traffic direction, rate control is always enabled or disabled for both traffic shaper instances of the selected port.

=>:eth switch shaper config shaper=0 unicast=enabled control=enabled unknown=enableddiscard=enabled=>:eth switch shaper config shaper=1 multicast=enabled broadcast=enabled discard=enabled=>:eth switch shaper iflistPort Direction Shaper State Speed BurstSize Flags

1 egress 0 off 0 0 nnnnnn

1 ingress 0 off 0 0 u--cad

1 ingress 1 off 0 0 -mb--d










State: configurable per port and shaper type (ingress or egress).Speed & BurstSize: configurable per port and shaper. Only limited set of values.

Input is rounded.Flags: configurable per ingress shaper: u(nicast), m(ulticast), b(roadcast),

c(ontrol), a(ddress unknown), d(iscard), n(ot applicable).

The same frametype can not be set on different ingress shapers.

Before you configure rate control on a switch port, disable the corresponding bridge port first.

E-DOC-CTC-20080229-0002 v1.0 13

14


Speed: this parameter specifies the speed limit, expressed in bits per second (bps). The value can only be changed in discrete steps:

Steps of 64 000 bps within the range from 64 000 bps through 1 792 000 bps.

The next possible value after 1 792 000 bps is 2 000 000 bps.

Steps of 1 000 000 bps within the range from 2 000 000 bps through 100 000 000 bps.

As only discrete steps are possible, rounding is used.

Burstsize: this parameter specifies the maximum burst size, expressed in kilobytes (KB). This value can only be changed in discrete steps within the range from 6 KB through 126 KB. The possible values are 6 KB, 10 KB, 18 KB, 34 KB, 66 KB and 126 KB. Only these values can be specified during configuration (no rounding is used).

Each value is mapped to the next (higher) possible discrete value. For example, 64 001 bps is mapped to 128 000 bps.

=>:eth bridge ifconfig intf=ethport2 portstate=disabled=>:eth switch shaper ifconfig port=2 ingress=disabled shaper=0 state=enabled speed=90000000burstsize=18=>:eth switch shaper ifconfig port=2 ingress=enabled shaper=0 state=enabled speed=30000000burstsize=10=>:eth switch shaper iflistPort Direction Shaper State Speed BurstSize Flags1 egress 0 off 0 0 nnnnnn1 ingress 0 off 0 0 u--cad1 ingress 1 off 0 0 -mb--d2 egress 0 on 90000000 18 nnnnnn2 ingress 0 on 30000000 10 u--cad2 ingress 1 on 100000000 126 -mb--d3 egress 0 off 0 0 nnnnnn3 ingress 0 off 0 0 u--cad3 ingress 1 off 0 0 -mb--d4 egress 0 off 0 0 nnnnnn4 ingress 0 off 0 0 u--cad4 ingress 1 off 0 0 -mb--dState: configurable per port and shaper type (ingress or egress).

Speed & BurstSize: configurable per port and shaper. Only limited set of values.

Input is rounded.

Flags: configurable per ingress shaper: u(nicast), m(ulticast), b(roadcast),c(ontrol), a(ddress unknown), d(iscard), n(ot applicable).The same frametype can not be set on different ingress shapers.

=>:eth bridge ifconfig intf=ethport2 portstate=forwarding

E-DOC-CTC-20080229-0002 v1.0


3.3 Storm Control

IntroductionA traffic storm occurs when frames flood the LAN, creating excessive traffic and degrading the network performance.

The storm control feature prevents the switch ports from being disrupted by a broadcast, multicast, or unicast traffic storm. Storm control is thus rather a security mechanism than a QoS feature.

The CLI commands described in this section belong to the command group :eth switch storm.

Listing the storm control configurationTo look up the storm control configuration, execute following command:

Configuring per port storm controlIngress storm control can be activated per port with the command :eth switch storm ifconfig.

The port must be specified:


Optionally, following parameters can be specified:

State: this parameter specifies whether storm control is enabled or not. By default, storm control is disabled.

Rate: this parameter specifies the maximum allowed rate of the selected traffic types, expressed in tenths of one percent (0.1%). Only specific values can be used during configuration (no rounding is used).

In case of a business device, the only possible values are 33, 50, 100 and 200.

In case of a residential device, the only possible values are 100, 200, 300 and 400.

Burstsize: this parameter specifies the maximum burst size, expressed in kilobytes (KB). This value can only be changed in discrete steps of 2 KB within the range from 2 KB through 8 KB. As only discrete steps are possible, rounding is used.

=>:eth switch storm iflistPort State Rate(0.1%) BurstSize Flags1 off 0 0 ---2 off 0 0 ---3 off 0 0 ---4 off 0 0 ---Rate & BurstSize: only limited set of values. Input is rounded.Flags: b(roadcast), m(ulticast), a(ddress unknown).

Each value is mapped to the next (higher) possible discrete value. For example, 3 KB is mapped to 4 KB.

E-DOC-CTC-20080229-0002 v1.0 15

16


Storm control can be used to control bursts of different traffic types. For each traffic type, the corresponding parameter can be enabled or disabled. By default, all traffic types are disabled.

Broadcast: broadcast traffic

Multicast: multicast traffic

Unknown: traffic that resulted in a destination look up failure (DLF). This means that the destination address is unknown.

=>:eth switch storm ifconfig port=1 state=enabled rate=33 burstsize=5 broadcast=enabled=>:eth switch storm iflistPort State Rate(0.1%) BurstSize Flags1 on 33 6 b--

2 off 0 0 ---3 off 0 0 ---4 off 0 0 ---Rate & BurstSize: only limited set of values. Input is rounded.Flags: b(roadcast), m(ulticast), a(ddress unknown).

Storm control can only be enabled if both the rate and burst size are different from zero and at least one traffic type is selected.

E-DOC-CTC-20080229-0002 v1.0

4| QoS Capabilities of the Ethernet Bridge

4 QoS Capabilities of the Ethernet Bridge

OverviewFollowing QoS features are described in this chapter:

Topic Page

“4.1 802.1p User Priority Regeneration” 18

“4.2 Ingress Classification” 20

“4.3 Queuing” 28

“4.4 Priority Tagging of Egress Traffic” 31

“4.5 802.1p User Priority Preservation” 32

E-DOC-CTC-20080229-0002 v1.0 17

18


4.1 802.1p User Priority Regeneration

IntroductionThe VLAN tag of VLAN-tagged frames contains a three-bit user priority field. This field indicates the 802.1p user priority, which is also referred to as the Priority Code Point (PCP) or the Class of Service (CoS).

802.1p user priority regeneration maps the 802.1p user priority value of a received frame to the regenerated

user priority, also referred to as the VLAN user priority. The mapping is based on the user priority regeneration table.

The regenerated user priority corresponds to a fixed internal class.

User priority regeneration tableThe user priority regeneration table maps the 802.1p user priority of a received frame to a regenerated user priority. The table contains eight entries, one entry for each possible value of the 802.1p user priority.

To look up the regeneration table of a bridge port, execute following command:

The regeneration table is indicated by a string. The default table is represented by the string 0 1 2 3 4 5 6 7. This must be interpreted as follows:

802.1p user priority 0 is mapped to the regenerated user priority 0.








Configuring the regeneration tableThe user priority regeneration table can be configured independently for each bridge port and for each value of the 802.1p user priority. For the configuration of an entry, the full range of values can be used for the regenerated user priority.

If you want to configure the regeneration as follows:


=>:eth bridge iflist intf=ethport1ethport1 : dest : ethif1

Connection State: connected Retry: 10Priority Tagging: NA (destination switch interface)Port: ethport1 PortNr: 1 PortState: forwarding Interface: upMulticast filter: disabled Dynamic VLAN : disabledWAN : disabledIGMP snooping : enabledTransparent Prio: disabledBPDU Filtering : disabledExtra Tagging : noneVLAN: Default VLAN: default Ingressfiltering: disabled Acceptvlanonly: disabledVLAN: Priority: disabled IP Prec: disabled Priority: 0Regeneration table: 0 1 2 3 4 5 6 7

RX bytes: 49421 frames: 14TX bytes: 10512 frames: 22 dropframes: 0

E-DOC-CTC-20080229-0002 v1.0









Then the regeneration table must be configured with the string 7 6 5 4 3 2 1 0.

To configure the regeneration table, execute following command:

From regenerated user priority to internal classEach regenerated user priority is mapped to one of the sixteen internal classes. This mapping is fixed. Following table shows this mapping:

=>:eth bridge ifconfig intf=ethport1 regenprio=76543210

Type the string without interspacing.

In case of a business device, the same user priority regeneration table is applied to all bridge ports that are connected to a switch port.

Regenerated user priority Internal class

7 15

6 14

13

12

11

5 10

9

4 8

7

3 6

5

0 (Default) 4 (Default)

3

2 2

1

1 0

E-DOC-CTC-20080229-0002 v1.0 19

20


4.2 Ingress Classification

IntroductionDuring ingress classification, an internal class is assigned to a received frame, based on a frame’s medium priority and evaluated priority:

The medium priority: this priority is based on the medium (physical interface) the received frame is entering on.

The evaluated priority: this priority is based on the information fields in the received frame. The used information can be:

802.1p user priority field

port default priority

IP precedence field

DSCP field

The resulting internal class assigned to the frame is also referred to as the final priority or internal priority.

E-DOC-CTC-20080229-0002 v1.0


4.2.1 Medium Priority

Medium priorityThe medium priority is an internal class. Hence, the value can range from 0 through 15.

Based on the medium (physical interface) the frame is entering on, a medium priority is assigned to each incoming frame:

For Ethernet, the medium priority is 4.

For ATM, the medium priority depends on the ATM QoS profile set on the interface (ATM PVC). An ATM QoS profile consists of a CTD (Connection Traffic Descriptor) for the upstream and downstream direction. For each CTD, an ATM QoS category was specified. Following table shows the mapping between the ATM QoS category and the medium priority:

ATM QoS category Internal class

CBR 15

VBR-rt 14

VBR-nrt (low CDVT) 13

GFR (low CDVT) 12

VBR-nrt (high CDVT) 11

GFR (high CDVT) 10

9

8

UBR BCS 7 7

UBR BCS 6, ABR 6

UBR BCS 5, UBR-mdcr 5

UBR BCS 4, UBR 4 (default)

UBR BCS 3 3

UBR BCS 2 2

UBR BCS 1 1

UBR BCS 0 0

E-DOC-CTC-20080229-0002 v1.0 21

22


4.2.2 Evaluated Priority

Evaluated priorityThe evaluated priority is an internal class. Hence, the value can range from 0 through 15.

Based on the information fields in a received frame, an evaluated priority is assigned to the frame. The information fields that can be used are:

The 802.1p user priority field in the VLAN tag of a VLAN-tagged frame.

The IP precedence field in the IP header of a packet.

The DSCP field in the IP header of a packet.

During configuration, one information field must be selected. The value of the selected field is translated to the evaluated priority.

Configuring the classification criterionOnly one information field can be used to determine the evaluated priority. The information fields can not be combined. The classification criterion indicates which information field must be used.

The parameter ipprec of the command :eth bridge ifconfig indicates the classification criterion. The value of this parameter can be set to:

Disabled: in this case, the 802.1p user priority is used as classification criterion. The 802.1p user priority is translated to the internal class that corresponds to the regenerated user priority.

Precedence: in this case, the IP precedence value is used as classification criterion. If no IP packet is found in the frame, the classification criterion “disabled” is used.

DSCP: in this case, the DSCP value is used as classification criterion. If no IP packet is found in the frame, the classification criterion “disabled” is used.

The parameter can be configured per bridge port. By default, the parameter is set to disabled. To change the selected classification criterion, modify the parameter ipprec as follows:

For more information on these information fields, see “ Classification-related frame fields” on page 5.

=>:eth bridge ifconfig intf=ethport1 ipprec=precedence=>:eth bridge iflist intf=ethport1ethport1 : dest : ethif1

Connection State: connected Retry: 10Priority Tagging: NA (destination switch interface)Port: ethport1 PortNr: 1 PortState: forwarding Interface: upMulticast filter: disabled Dynamic VLAN : disabledWAN : disabledIGMP snooping : enabledTransparent Prio: disabledBPDU Filtering : disabledExtra Tagging : noneVLAN: Default VLAN: default Ingressfiltering: disabled Acceptvlanonly: disabledVLAN: Priority: overwrite IP Prec: precedence Priority: 2Regeneration table: 0 1 2 3 4 5 6 7RX bytes: 49421 frames: 14TX bytes: 10512 frames: 22 dropframes: 0

When the 802.1p user priority is selected as classification criterion, frames received by the Thomson Gateway will have a VLAN header. It is required to configure explicitly that the bridge is not VLAN aware. If the bridge is VLAN aware, you must define all the VLANs that have to pass through the Thomson Gateway. Otherwise, unknown VLANs will be dropped.

E-DOC-CTC-20080229-0002 v1.0


Configuring the default user priorityIf the 802.1p user priority field is selected as classification criterion and the received frame is untagged, a default user priority is used. This default user priority can be configured per bridge port.

By default, the default user priority is set to 0. To change the default user priority of a bridge port, execute following command:

IP precedence mapping tableThe IP precedence mapping table defines the translation from the IP precedence value to an internal class. The following table shows this mapping:

=>:eth bridge ifconfig intf=ethport1 priority=2=>:eth bridge iflist intf=ethport1ethport1 : dest : ethif1

Connection State: connected Retry: 10Priority Tagging: NA (destination switch interface)Port: ethport1 PortNr: 1 PortState: forwarding Interface: upMulticast filter: disabled Dynamic VLAN : disabledWAN : disabledIGMP snooping : enabledTransparent Prio: disabledBPDU Filtering : disabledExtra Tagging : noneVLAN: Default VLAN: default Ingressfiltering: disabled Acceptvlanonly: disabledVLAN: Priority: overwrite IP Prec: precedence Priority: 2

Regeneration table: 0 1 2 3 4 5 6 7RX bytes: 49421 frames: 14TX bytes: 10512 frames: 22 dropframes: 0

The user priority regeneration table is only used to remap the 802.1p user priority of a VLAN-tagged frame. The table is not used to remap the default user priority assigned to untagged frames.

IP precedence (label) Internal class

6 (Internetwork Control), 7 (Network Control) 15

5 (CRITIC/ECP) 14

4 (Flash Override) 13

12

3 (Flash) 11

10

2 (Immediate) 9

8

1 (Priority) 7

6

5

0 (Routine) 4 (default)

3

2

E-DOC-CTC-20080229-0002 v1.0 23

24


Configuring the IP precedence mapping tableThe mapping from the IP precedence value to an internal class is not fixed. The mapping table can be configured for the whole bridge (not per bridge port) using the parameter precedencemap of the command :eth bridge config:

DSCP mapping tableSome DSCP values have been standardized by the IETF. The translation from the DSCP value to an internal class is defined by the DSCP mapping table, which is depicted below (EF = Expedited Forwarding, CS = Class Selector, AF = Assured Forwarding):

1

0

IP precedence (label) Internal class

=>:eth bridge configAgeing : 300Filter : no_WAN_broadcastVLAN : disabledIPQoS precedence map for TOS:

IP priority QoS internal class

0 4

1 7

2 9

3 11

4 13

5 14

6 15

7 15

=>:eth bridge config precedencemap=1,3,5,7,9,11,13,15=>:eth bridge configAgeing : 300Filter : no_WAN_broadcastVLAN : disabledIPQoS precedence map for TOS:

IP priority QoS internal class0 11 32 53 74 95 116 137 15

The precedence map is a string that consists of exactly eight internal classes, separated by commas.

DSCP decimal value (name) Internal class

48 (CS6), 56 (CS7) 15

40 (CS5), 46 (EF) 14

32 (CS4), 34 (AF41) 13

E-DOC-CTC-20080229-0002 v1.0


36 (AF42), 38 (AF43) 12

24 (CS3), 26 (AF31) 11

28 (AF32), 30 (AF33) 10

16 (CS2), 18 (AF21) 9

20 (AF22), 22 (AF23) 8

8 (CS1), 10 (AF11) 7

12 (AF12), 14 (AF13) 6

5

0 (CS0, Best-effort) 4 (default)

3

2

1

0

You can not configure the DSCP mapping table.

DSCP decimal value (name) Internal class

E-DOC-CTC-20080229-0002 v1.0 25

26


4.2.3 Internal Priority

Internal priorityThe internal priority is an internal class. Hence, the value can range from 0 through 15.

This priority is the final priority assigned to a received frame, based on a frame’s medium priority and evaluated priority.

Configuring the internal priority mechanismThe mechanism to determine the internal priority is configured by the parameter prioconfig, which has one of the following values:

Disabled: in this case, the evaluated priority is not used, the medium priority is assigned to the internal priority.

Overwrite: in this case, the medium priority is not used, the evaluated priority is assigned to the internal priority.

Increase: in this case, the medium priority is compared with the evaluated priority:

If the evaluated priority is better (higher value), the internal priority is set to the evaluated priority.

In all other cases, the internal priority is simply the medium priority.

The parameter prioconfig of the command :eth bridge ifconfig can be configured for each bridge port. By default, the parameter is disabled. To modify the configuration, execute following command:

Summary: priority mapping tableSeveral mapping tables were described in the previous subsections. The following table summarizes these mapping tables:

=>:eth bridge ifconfig intf=ethport1 prioconfig=overwrite=>:eth bridge iflist intf=ethport1ethport1 : dest : ethif1

Connection State: connected Retry: 10Priority Tagging: NA (destination switch interface)Port: ethport1 PortNr: 1 PortState: forwarding Interface: upMulticast filter: disabled Dynamic VLAN : disabledWAN : disabledIGMP snooping : enabledTransparent Prio: disabledBPDU Filtering : disabledExtra Tagging : noneVLAN: Default VLAN: default Ingressfiltering: disabled Acceptvlanonly: disabledVLAN: Priority: overwrite IP Prec: precedence Priority: 2Regeneration table: 0 1 2 3 4 5 6 7RX bytes: 49421 frames: 14TX bytes: 10512 frames: 22 dropframes: 0

Regenerated

user priority

IP precedence

(ToS)

DSCP

(DiffServ)


7 6, 7 CS6, CS7 CBR 15

6 5 CS5, EF VBR-rt 14

4 CS4, AF41 VBR-nrt (low CDVT) 13

E-DOC-CTC-20080229-0002 v1.0


AF42, AF43 GFR (low CDVT) 12

3 CS3, AF31 VBR-nrt (high CDVT) 11

5 AF32, AF33 GFR (high CDVT) 10

2 CS2, AF21 9

4 AF22, AF23 8

1 CS1, AF11 UBR BCS 7 7

3 AF12, AF13 UBR BCS 6, ABR 6

UBR BCS 5, UBR-mdcr 5

0 0 BE, CS0 UBR BCS 4, UBR 4 (default)

UBR BCS 3 3

2 UBR BCS 2 2

UBR BCS 1 1

1 UBR BCS 0 0

Regenerated

user priority

IP precedence

(ToS)

DSCP

(DiffServ)


E-DOC-CTC-20080229-0002 v1.0 27

28


4.3 Queuing

IntroductionThe use of Ethernet QoS on the Thomson Gateway only improves the performance of egress traffic if the egress interfaces are QoS enabled.

Based on the assigned internal class, a frame can be queued into the correct queue at the egress side. In this subsection, we briefly describe two types of egress interfaces:

Wireless interface

ATM interface (single PVC)

The internal class can also be used to map the frames to the correct ATM PVC. To this end, following concept is implemented:

ATM interface bundle (multiple PVCs)

QoS on the wireless interfaceIn case of a wireless interface, WMM (Wi-Fi Multi Media) supports QoS:

First, the internal class assigned to a frame is mapped to the corresponding 802.1p user priority.

Next, the 802.1p user priority is mapped to a WMM AC (Access Category), which corresponds to a queue.

To enable WMM QoS support on the wireless interface, execute following command:

The class-to-queue mapping is shown in following table. The four access categories correspond to four queues.

=>:wireless qos config mode=wmm

Internal class VLAN user priority Access category (AC)

15 7 AC_VO (Voice)

14 6 AC_VO (Voice)

13

12

11

10 5 AC_VI (Video)

9

8 4 AC_VI (Video)

7

6 3 AC_BE (Best-effort)

5

4 0 AC_BE (Best-effort)

3

2 2 AC_BK (Background)

E-DOC-CTC-20080229-0002 v1.0


QoS on an ATM interfaceIP QoS can be enabled per ATM interface, thus per PVC.

If IP QoS is disabled, no IP QoS queues will be instantiated. As a result, all data (either classified or unclassified) is sent to a fixed destination queue (the AAL5 segmentation queue) for that PVC.

If IP QoS is enabled, all six IP QoS queues are instantiated on the PVC. A packet queuing mechanism, which is active on ATM interface level is enabled. This queuing mechanism handles incoming packets according to their internal class. If no internal class is assigned to the data, the data is forwarded to the lowest priority best-effort queue.

To enable IP QoS on an ATM interface with destination phonebookentry, execute following command:

The class-to-queue mapping is shown in following table. By default, the first queue (EF) has strict priority over the four queues (AF) with WFQ scheduling. These four queues have strict priority over the best-effort queue.

1

0 1 AC_BK (Background)

You cannot configure the class-to-queue mapping, which is fixed.

Internal class VLAN user priority Access category (AC)

=>:ipqos config dest=phonebookentry state=enabled

Internal class QoS queue (traffic type)

15 5 (Real-time EF data)

14

13 4 (AF Class 4)

12

11 3 (AF Class 3)

10

9 2 (AF Class 2)

8

7 1 (AF Class 1)

6

E-DOC-CTC-20080229-0002 v1.0 29

30


ATM interface bundlesATM interface bundles can be used to forward frames over multiple PVCs based on the internal class assigned to each frame.

If the different PVCs would be connected to the bridge via different bridge ports, the bridge learning process would learn the same remote MAC address from different bridge ports. To prevent this, the PVCs are bundled into an ATM interface bundle, which is connected to the bridge via a single bridge port. Up to 6 PVCs can be added to a single ATM interface bundle.

To forward frames over multiple PVCs based on their internal class, the used ATM interface bundle policy is the priority mapping policy. Within the priority mapping policy, every PVC that is a member of the ATM interface bundle can have a priority range configured (as low and high selector values). All frames with a priority within the first matching member priority range will be forwarded to this PVC.

5 0 (Best-effort)

4 (default)

3

2

1

0

You cannot configure the class-to-queue mapping, which is fixed.

Internal class QoS queue (traffic type)

The ATM interface bundles feature is only available on selected products.

E-DOC-CTC-20080229-0002 v1.0


4.4 Priority Tagging of Egress Traffic

What is priority tagging?Priority tagging means that a VLAN tag is added to a frame with the 802.1p user priority field set to a given value and the VLAN ID equal to 0. The value that is assigned to the 802.1p user priority field is based on the

internal class of the frame.

This feature is intended to be used by:

Bridge ports located at the WAN side.

Bridge ports that are untagged member of a VLAN on the bridge.

Configuring priority taggingThe priority tagging feature can be configured per bridge port.

By default, priority tagging is disabled. To enable the feature, execute following command:

Listing informationTo check the state of the priority tagging feature, execute following command:

=>:eth bridge ifconfig intf=eth_data priotag=enabled

The priority tagging feature is not available (NA) for bridge ports that are connected to a switch port.

=>:eth bridge iflist intf=eth_dataeth_data : dest : atm_0.38

Connection State: connected Retry: 10Priority Tagging: Enabled

Port: wan0 PortNr: 6 PortState: forwarding Interface: upMulticast filter: disabled Dynamic VLAN : disabledWAN : enabledIGMP snooping : enabledTransparent Prio: disabledBPDU Filtering : disabledExtra Tagging : noneVLAN: Default VLAN: default Ingressfiltering: disabled Acceptvlanonly: disabledVLAN: Priority: disabled IP Prec: disabled Priority: 0Regeneration table: 0 1 2 3 4 5 6 7RX bytes: 49421 frames: 14TX bytes: 10512 frames: 22 dropframes: 0

E-DOC-CTC-20080229-0002 v1.0 31

32


4.5 802.1p User Priority Preservation

What is 802.1p user priority preservation?At the egress side, each bridge port can be configured with an 802.1p user priority preservation flag:

If user priority preservation is enabled, the bridge port tags each frame with the original 802.1p user

priority as present in that frame when it was received at the ingress side. Therefore, the priority preservation feature is also referred to as the priority transparency feature.

If user priority preservation is disabled, the priority in each tagged frame is based on the internal class of the frame.

This feature is intended to be used by:

Bridge ports located at the WAN side.

Bridge ports that are tagged member of a VLAN on the bridge.

Configuring 802.1p user priority preservationThe user priority preservation feature can be configured per bridge port.

By default, user priority preservation is disabled. To enable the feature, execute following command:

Listing informationTo check the state of the user priority preservation feature, execute following command:

Applications802.1p user priority preservation enables classification without changing the original 802.1p user priority:

On the one hand, an incoming frame is assigned to an internal class, for example based on the value of the DSCP field. This internal class is used to queue the frame in the correct queue.

On the other hand, the original 802.1p user priority of the (tagged) frame is not overwritten. If priority tagging would be based on the internal class assigned to the frame during classification, the original 802.1p user priority may be overwritten.

=>:eth bridge ifconfig intf=eth_data priotransparent=enabled

=>:eth bridge iflist intf=eth_dataeth_data : dest : atm_0.38

Connection State: connected Retry: 10Priority Tagging: DisabledPort: wan0 PortNr: 6 PortState: forwarding Interface: upMulticast filter: disabled Dynamic VLAN : disabledWAN : enabledIGMP snooping : enabledTransparent Prio: enabled

BPDU Filtering : disabledExtra Tagging : noneVLAN: Default VLAN: default Ingressfiltering: disabled Acceptvlanonly: disabledVLAN: Priority: disabled IP Prec: disabled Priority: 0Regeneration table: 0 1 2 3 4 5 6 7RX bytes: 49421 frames: 14TX bytes: 10512 frames: 22 dropframes: 0

E-DOC-CTC-20080229-0002 v1.0

5| Logical Ethernet Interfaces and QoS

5 Logical Ethernet Interfaces and QoS

IntroductionLogical Ethernet interfaces can be configured to support following Ethernet QoS feature:

Priority tagging of traffic at the WAN-side

Priority tagging parameterPriority tagging by a logical Ethernet interface can be enabled using the priotag parameter of the :eth ifconfig command.

If the parameter is disabled, priority tagging is not supported. By default, priority tagging of egress traffic is disabled.

If the parameter is enabled, priority tagging is supported.

Configuring WAN-side priority taggingTo create a logical Ethernet interface that transmits priority-tagged frames, execute following steps:

1 Create a logical Ethernet interface:

2 Configure the logical Ethernet interface to support WAN-side priority tagging:

3 Activate the logical Ethernet interface:

=>:eth ifconfig intf=eth_data priotag=enabled=>:eth iflist intf=eth_dataeth_data : Dest: atm_data

Connection State: not-connected Retry: 10WAN: EnabledPriority Tagging: Enabled

PortNr: (unassigned)VLAN: default

The priority tagging feature of logical Ethernet interfaces is very similar to the priority tagging feature of bridge ports. For more information on the priority tagging feature of bridge ports, see “4.4 Priority Tagging of Egress Traffic” on page 31.

=>:eth ifadd intf=eth_data

=>:eth ifconfig intf=eth_data dest=atm_data wan=enabled priotag=enabled

=>:eth ifattach intf=eth_data

E-DOC-CTC-20080229-0002 v1.0 33

34

5| Logical Ethernet Interfaces and QoS

E-DOC-CTC-20080229-0002 v1.0

6| Layer 2 QoS and Layer 3 QoS Interaction

6 Layer 2 QoS and Layer 3 QoS Interaction

IntroductionAt the ingress side, incoming frames are classified at Layer 2, as described in this document. This means that an internal class is assigned to each received Ethernet frame, thus also to the IP packet in the payload of the frame. Following questions arise with regard to this assigned internal class:

What is the internal class of a packet that also passes through the IP router of the Thomson Gateway?

What is the internal class of a packet that is originated by the Thomson Gateway?

At the egress side, WAN-side priority tagging can be used at Layer 2. Following question arises:

Is Layer 3 QoS influenced by Layer 2 QoS and vice versa?

This chapter shortly answers these questions.

Internal class after Layer 3 classificationWhen a packet passes through the IP router of the Thomson Gateway, the packet is classified again. This is done in three steps:

1 The Layer 3 classification assigns a label to each packet. This label is only of internal significance and can be used in packet forwarding (routing label) and QoS definition (QoS label).

2 The label is mapped to the packet class associated with it, configured by the defclass parameter of the :label modify command:

0...15: in this case, the parameter value indicates the packet class.

Dscp: if this value is used, the packet class is set to the internal class corresponding with the value of the DSCP field of the packet.

Default: if the parameter value is set to default, the Thomson Gateway’s default value 4 is used.

3 The packet class is mapped to the Layer 3 internal class. By configuring the classification parameter of the :label modify command, it is determined how the Layer 2 internal class is taken into account:

Ignore: in this case, the assigned packet class is ignored. The internal class assigned by Layer 2 is not overwritten.

Overwrite: in this case, the internal class assigned by Layer 2 is ignored. The internal class is set to the packet class.

Increase: in this case, the label classification only sets the internal class if the packet class value is higher than the internal class value assigned by Layer 2.

Internal class of Thomson Gateway originated packetsPackets that are originated by the Thomson Gateway itself also receive a label. The corresponding packet class of the label is the internal class assigned to the packet.

The system service manager can be used to explicitly assign a QoS label to an application. If QoS label assignment is used for an application, no label classification is done.

Application helpers can be used to establish signalling (parent)/ data (child) relationships. Label inheritance means that when parent connections get labels associated, these labels are also automatically assigned to the child connections. It is also possible to define separate child QoS labels.

E-DOC-CTC-20080229-0002 v1.0 35

36

6| Layer 2 QoS and Layer 3 QoS Interaction

IP marking and priority taggingWhile Layer 2 supports priority tagging, IP marking is used at Layer 3:

IP marking: at Layer 3, IP marking can be enabled per label using the tosmarking parameter of the :label modify command. IP marking can be used to set the IP ToS byte, IP precedence field or DSCP field of egress packets, regardless the internal class of the packet.

Priority tagging: at Layer 2, priority tagging can be used.The 802.1p user priority is based on the internal class of the egress frames. Depending on the configuration, this internal class can be influenced by the Layer 3 classification.

E-DOC-CTC-20080229-0002 v1.0

Ethernet QoS Configuration Guide - Avanzada 7 · 2 E-DOC-CTC-20080229-0002 v1.0 About this Ethernet...

Documents

Transcript of Ethernet QoS Configuration Guide - Avanzada 7 · 2 E-DOC-CTC-20080229-0002 v1.0 About this Ethernet...